LSTMModel class

Description

LSTMModel class for further use in predict function

Usage

LSTMModel(
  lstm_model,
  scale_output,
  scaler_output,
  scale_input,
  scaler_input,
  tsLag,
  xregLag,
  model_structure,
  batch_size,
  lags_as_sequences,
  stateful
)

Arguments

Value

LSTMModel object

References

Paul, R.K. and Garai, S. (2021). Performance comparison of wavelets-based machine learning technique for forecasting agricultural commodity prices, Soft Computing, 25(20), 12857-12873

Examples

if (keras::is_keras_available()){
  y<-rnorm(100,mean=100,sd=50)
  x1<-rnorm(100,mean=50,sd=50)
  x2<-rnorm(100, mean=50, sd=25)
  x<-cbind(x1,x2)
  TSLSTM<-ts.lstm(ts=y,
                  xreg = x,
                  tsLag=2,
                  xregLag = 0,
                  LSTMUnits=5,
                  ScaleInput = 'scale',
                  ScaleOutput = 'scale',
                  Epochs=2)
}

Create lead/lags of a variable

Description

Create an array with lead/lags of an input variable.

Usage

lagmatrix(x, lag)

Arguments

Value

An array with the resulting leads and lags (columns).

Note

This code was copied from the ts.utils package to avoid the archive operations of the smooth package in 16-02-2025. This function might be deprecated in future releases to use the one from ts.utils again.

Author(s)

Nikolaos Kourentzes, nikolaos@kourentzes.com

Examples

  x <- rnorm(10)
  lagmatrix(x,c(0,1,-1))

Min-Max Scaling of a Matrix

Description

This function applies min-max scaling to a matrix. Each column of the matrix is scaled independently. The scaling process transforms the values in each column to a specified range, typically [0, 1]. The function subtracts the minimum value of each column (if 'min' is 'TRUE' or a numeric vector) and then divides by the range of each column (if 'range' is 'TRUE' or a numeric vector).

Usage

minmax_scale(x, min = TRUE, range = TRUE)

Arguments

Value

A matrix with the same dimensions as 'x', where each column has been scaled according to the min-max scaling process.

Examples

  data <- matrix(rnorm(100), ncol = 10)
  scaled_data <- minmax_scale(data)

Predict using a Trained LSTM Model

Description

This function makes predictions using a trained LSTM model for time series forecasting. It performs iterative predictions where each step uses the prediction from the previous step. The function takes into account the lags in both the time series data and the exogenous variables.

Usage

## S3 method for class 'LSTMModel'
predict(
  object,
  ts,
  xreg = NULL,
  xreg.new = NULL,
  horizon = NULL,
  BatchSize = NULL,
  ...
)

Arguments

Value

A vector containing the forecasted values for the specified horizon.

Examples

  if (keras::is_keras_available()){
      y<-rnorm(100,mean=100,sd=50)
      x1<-rnorm(150,mean=50,sd=50)
      x2<-rnorm(150, mean=50, sd=25)
      x<-cbind(x1,x2)
      x.tr <- x[1:100,]
      x.ts <- x[101:150,]
      TSLSTM<-ts.lstm(ts=y,
                      xreg = x.tr,
                      tsLag=2,
                      xregLag = 0,
                      LSTMUnits=5,
                      ScaleInput = 'scale',
                      ScaleOutput = 'scale',
                      Epochs=2)
      current_values <- predict(TSLSTM, xreg = x.tr, ts = y)
      future_values <- predict(TSLSTM, horizon=50, xreg = x, ts = y, xreg.new = x.ts)
   }

Summary of a Trained LSTM Model

Description

This function generates the summary of the LSTM model.

Usage

## S3 method for class 'LSTMModel'
summary(object, ...)

Arguments

Value

A vector containing the forecasted values for the specified horizon.

Examples

  if (keras::is_keras_available()){
      y<-rnorm(100,mean=100,sd=50)
      x1<-rnorm(100,mean=50,sd=50)
      x2<-rnorm(100, mean=50, sd=25)
      x<-cbind(x1,x2)
      TSLSTM<-ts.lstm(ts=y,
                      xreg = x,
                      tsLag=2,
                      xregLag = 0,
                      LSTMUnits=5,
                      ScaleInput = 'scale',
                      ScaleOutput = 'scale',
                      Epochs=2)
      # Assuming TSLSTM is an LSTMModel object created using ts.lstm function
      summary(TSLSTM)
  }

Long Short Term Memory (LSTM) Model for Time Series Forecasting

Description

The LSTM (Long Short-Term Memory) model is a Recurrent Neural Network (RNN) based architecture that is widely used for time series forecasting. Min-Max transformation has been used for data preparation. Here, we have used one LSTM layer as a simple LSTM model and a Dense layer is used as the output layer. Then, compile the model using the loss function, optimizer and metrics. This package is based on 'keras' and TensorFlow modules.

Usage

ts.lstm(
  ts,
  xreg = NULL,
  tsLag = NULL,
  xregLag = 0,
  LSTMUnits,
  DenseUnits = NULL,
  DropoutRate = 0,
  Epochs = 10,
  CompLoss = "mse",
  CompMetrics = "mae",
  Optimizer = optimizer_rmsprop,
  ScaleOutput = c(NULL, "scale", "minmax"),
  ScaleInput = c(NULL, "scale", "minmax"),
  BatchSize = 1,
  LSTMActivationFn = "tanh",
  LSTMRecurrentActivationFn = "sigmoid",
  DenseActivationFn = "relu",
  ValidationSplit = 0.1,
  verbose = 2,
  RandomState = NULL,
  EarlyStopping = callback_early_stopping(monitor = "val_loss", min_delta = 0, patience =
    3, verbose = 0, mode = "auto"),
  LagsAsSequences = TRUE,
  Stateful = FALSE,
  ...
)

Arguments

Value

LSTMmodel object

References

Paul, R.K. and Garai, S. (2021). Performance comparison of wavelets-based machine learning technique for forecasting agricultural commodity prices, Soft Computing, 25(20), 12857-12873

Examples

  if (keras::is_keras_available()){
      y<-rnorm(100,mean=100,sd=50)
      x1<-rnorm(100,mean=50,sd=50)
      x2<-rnorm(100, mean=50, sd=25)
      x<-cbind(x1,x2)
      TSLSTM<-ts.lstm(ts=y,
                      xreg = x,
                      tsLag=2,
                      xregLag = 0,
                      LSTMUnits=5,
                      ScaleInput = 'scale',
                      ScaleOutput = 'scale',
                      Epochs=2)
  }

Prepare data for Long Short Term Memory (LSTM) Model for Time Series Forecasting

Description

The LSTM (Long Short-Term Memory) model is a Recurrent Neural Network (RNN) based architecture that is widely used for time series forecasting. Min-Max transformation has been used for data preparation. Here, we have used one LSTM layer as a simple LSTM model and a Dense layer is used as the output layer. Then, compile the model using the loss function, optimizer and metrics. This package is based on 'keras' and TensorFlow modules.

Usage

ts.prepare.data(ts, xreg = NULL, tsLag, xregLag = 0)

Arguments

Value

dataset with all lags created from exogenous and time series data.

Examples

  y <- rnorm(100,mean=100,sd=50)
  x1 <- rnorm(100,mean=50,sd=50)
  x2 <- rnorm(100, mean=50, sd=25)
  x <- cbind(x1,x2)
  ts.prepare.data(y, x, 2, 4)